Risks and benefits of phase I trials: Eighteen-year experience from a single institution.

Abstract

e18146 Background: Phase I trials are a critical component of drug development; yet are considered non-therapeutic. This leads to recruitment barriers due to provider and patient discomfort. The largest meta-analysis of phase I studies has previously shown a response rate (RR) of 10.6%. Herein we report the results from our institution. Methods: Records of patients enrolled on phase I trials at our institution from January 1999 to December 2016 were reviewed. Recorded data included adverse events (AE); treatment related responses and deaths. Kaplan-Meier analysis and t-test were performed on the reviewed data. Results: During this period 774 patients were accrued on 64 phase I trials [43.8% cytotoxic, 45.3% biological, 6.2% both and 4.7% viral agents]. Primary cancer diagnoses included colorectal (25.2%), ovarian (17.6%), lung (7.8%), uterine (6.6%) and breast (6.1%). In total, 609 patients were evaluable for response, 41.1% had stable disease (SD) and overall RR was 7.7% (complete RR = 1.0%). Patients with overall clinical benefit (SD+response) had lower mean baseline WBC (4.83 vs 5.94 k/uL, p = 0.0008), ANC (2.92 vs 4.12 k/uL, p = 0.00007), platelets (209 vs 246 k/uL, p = 0.00007), LDH (280 vs 346 U/L, p = 0.0055) and higher serum albumin (3.97 vs 3.86 g/dL, p = 0.011) as compared to patients with progressive disease. Grade 3/4 non-hematological and hematological AE were seen in 28.5% and 19.9% patients, respectively. Treatment-related mortality was 0.8%. Patients with baseline LDH below the median (247 U/L) for the cohort had a higher median survival (312 days vs 201 days, p < 0.001, HR 0.61 95% CI 0.51-0.73). The median and mean duration on study were 56 and 87 days, respectively. Conclusions: This is one of the largest single-institution series of phase I oncology trials. Our RR of 7.7% [95% CI 5.9-10.1%] falls within the 95% CI of the RR of a majority of third line (and greater) chemotherapy regimens for solid tumors. Thus, the concept of 'non-therapeutic' nature of phase I studies needs reconsideration.